• Form of regression analysis that • Widely used technique for
seeks to predict future forecasting for its simplicity and movements along the seemingly robustness random walk taken • Holt-Winters’ multiplicative • take into account method is appropriate for series trends, seasonality, cycles, errors with seasonal variations and non-stationary aspects of a data set when making forecasts. Hourly demand (MW) of electricity in the month of May 2006
Hourly demand (MW) of electricity in the
month of December 2006 Forecasts based on entire month data of May 2006 for estimation
Forecasts based on last
week data of May 2006 for estimation
Forecasts based on entire
month data of December 2006 for estimation
Forecasts based on last
week data of Decemeber 2006 for estimation Estimation RMSE MAE MAPE Model period 776.34 610.55 2.2 ARIMA May 2006 759.92 611.33 2.18 ES 579.4 467.84 1.68 Last week of ARIMA May2006 707.58 565.32 2.04 ES 425.76 343.72 1.41 ARIMA December 2006 454.7 379.14 1.55 ES 375.23 281.35 1.15 Last week of ARIMA December 2006 379.76 327.84 1.34 ES